Applications Of SOLIDWORKS Flow Simulation Computational Fluid Dynamics .

APPLICATIONS OF SOLIDWORKS FLOW SIMULATION COMPUTATIONAL FLUID DYNAMICSSOFTWARE TO THE INVESTIGATION OF FIRESBy: John B. Holecek, P.E., CFEIPage 5combustible material next to a heated vent stack could reach an ignitabletemperature after the appliance started, then a time dependent analysiswould be used. In either case, “goals” should be specified. These goals areused by Flow Simulation as a convergence criterion in its iterative solutionprocess, although Flow Simulation has additional internal convergencecriteria. For example, in the above indicated example, specifying themaximum temperature of the combustible material next to the stack as a goalwould tell Flow Simulation to run the simulation until the combustible materialtemperature reached a maximum steady state. At that point, the simulationwould stop, and output results would be given.Flow Simulation also has several optional inputs that are available to allowbuilding a model that is well matched to the actual physical arrangement.Some inputs that are available are sources of heat, porous media and fans.Heating sources can be specified as surfaces that have a specified heat fluxor volumes that have specified heat generation rates. Constant temperaturesources can also be specified. Heat sources can also be made to vary overtime if a transient analysis is being conducted. This feature could be used tomodel the effect of a specified heat release rate from an object in a study offire spread. Porous media is a feature that treats a solid as non-homogenous,like a filter would be. This can be used to model flows in useful ways as willbe shown. Fans are a feature that models the effect of an actual fan, althoughthe geometry of a specific fan does not have to be present. An object isinserted in the model in a proper location and a flow versus pressure dropspecification made for the object. When ran, the “fan” will move the specifiedflow through the object against the calculated pressure drop.Limitations of SOLIDWORKS Flow SimulationAlthough SOLIDWORKS Flow Simulation will do a great deal, it is importantto know that there are certain things that it will not directly model. Thisincludes directly modeling chemically reacting flows like combustion, althoughit will model heat generation in a solid or fluid which can be used to model theenergy input from a combustion process. It will model mixed gas flows suchas a fuel gas in air, allowing modeling of fuel / air ratios under differingconditions. This capability has clear application to modeling the flow offugitive fuel gas that may have preceded a gas explosion. Flow Simulationwill also not directly model radiant exchange between a gas (including ionizedgases like a flame) and a solid to include attenuation of radiant exchange by

APPLICATIONS OF SOLIDWORKS FLOW SIMULATION COMPUTATIONAL FLUID DYNAMICSSOFTWARE TO THE INVESTIGATION OF FIRESBy: John B. Holecek, P.E., CFEIPage 6an absorbing medium such as water vapor. However, it will model radiantexchange between solids, and a flame could potentially be modeled as aradiating solid at a certain internal heat generation rate or temperature. Thepractical effect of these limitations varies depending on the problem underanalysis. The best way to show the application of the software is byreference to an example.Flow Simulation Example 1: A Burner in a Wooden Box, Modeling a GasBurnerOne question that arises in fire investigation is if a heat producing appliance(e.g. furnace, gas fireplace) created sufficiently high temperatures in adjacentcombustibles to ignite a fire. Frequently such appliances use fuel gas firedburners. To refine the usage of SOLIDWORKS Flow Simulation to examiningsuch questions, a test apparatus with gas burner was constructed andoutfitted with multiple temperature sensors. The goal of the testing was todetermine if Flow Simulation could be used to accurately simulate heatingfrom a gas burner despite Flow Simulation not directly modeling combustion.The apparatus, shown in figure 1, consisted of a propane fueled tube burnermounted inside a plywood box. The burner was equipped with a #56 orificeand supplied at 2.6-2.7kPa (10.5-11.0 inwc) pressure propane gas.Openings were included near floor level and in the top of the box.The CFD model was ran as an external analysis which includes theenvironment around the box. Flow was allowed into and out of the openingsof the box under the influence of the buoyant effects of gravity due to theburner heat. Conduction within solids was allowed as was convection to thewalls. Convective heat transfer coefficients were calculated using themethodology of Section 7-9 of JP Holman’s Heat Transfer text.3 Theseranged from 2.1-5.8 W/m2K (0.37-1.02 BTU/HrFT2F) for the differing surfaces.External air temperatures were set at 26.7 C (80 F), which was thetemperature in the lab. Radiation was not included in this model, as therelatively “clean” flame had a comparatively low proportion of energy emittedas radiant heat. Additionally, essentially all the energy released as radiantheat would transfer to the box interior surfaces as the openings in the boxwere small.As shown in figure 2, the burner was modeled by creating a solidrepresentation of the “flame” in the approximate size of the actual flame. The

APPLICATIONS OF SOLIDWORKS FLOW SIMULATION COMPUTATIONAL FLUID DYNAMICSSOFTWARE TO THE INVESTIGATION OF FIRESBy: John B. Holecek, P.E., CFEIPage 7solid “flame” was made porous and given an internal heat generation rateequal to the associated lower heating value of propane for the measuredflowrate of fuel. The flame was connected or “mated” to a model of the tubeburner which included the burner ports. Normally, a discharge of LP gas fromthe burner orifice entrains air into the burner’s venturi and the combinedmixture discharged from the ports. To simulate this process, a “fan” wasused to model flow through the burner. In Flow Simulation, “fans” arefeatures of a model that act as a real fan does, moving volume and/or massat a defined rate. In this case, a fan was added that moved a mass of airequal to the stoichiometrically correct volume for the gas flow measuredduring the test.As shown in figures 3-5, the model had a reasonably close agreement withmeasured temperatures and velocities. For example, referencing figure 3predicted versus measured exiting air temperatures at the top slot rangedfrom essentially exact agreement to a maximum difference of 21 C (121 Cversus 142 C). This is a range of 0-5.3% calculated as the differencebetween actual and predicted values over the value of the actual temperaturein absolute units (degrees Kelvin). Referencing figure 4, predicted versusmeasured inlet velocity varied from essential agreement to a maximumdifference of 0.13 m/s (.59 to .72 m/s). This is a range of 0-22% differencecalculated as the velocity difference divided by the actual value. Referencingfigure 5, the single interior top surface that was monitored averaged 112 Cwhile the predicted temperature was 113 C on one side of the slot and 98 Con the other side, a difference of 0% to 3.6% depending on what side isconsidered and calculated as the difference between actual and predictedvalues over the value of the actual temperature in absolute units (degreesKelvin).

APPLICATIONS OF SOLIDWORKS FLOW SIMULATION COMPUTATIONAL FLUID DYNAMICSSOFTWARE TO THE INVESTIGATION OF FIRESBy: John B. Holecek, P.E., CFEIPage 8Figure 1 – Flow Modeling Test Apparatus of a Wooden Box with an Internal Burner.The Corresponding Solid Model is Shown to the Right.“Flame”, Modeled as aPorous Material with aSpecified Heat Generationrate. The porous mediaallows airflow from the “fan”to discharge through the“flame”SOLIDWORKS “Fan” insideburner tube specified to havemass flowrate equal to thestoichiometrically correctquantity of airflow for thequantity of gas burned.Figure 2 – Solid Model of the Burner Showing Several Features and the Airflow Dueto the “Fan” Feature.

APPLICATIONS OF SOLIDWORKS FLOW SIMULATION COMPUTATIONAL FLUID DYNAMICSSOFTWARE TO THE INVESTIGATION OF FIRESBy: John B. Holecek, P.E., CFEIPage 9Figure 3 – A graph of air temperature exiting the lengthwise centerline of theopening in the top of the box shown in figure 1. The line is a plot of temperaturesalong this centerline as predicted by the Simulation. The individual points aretemperatures measured during the actual test trials.Figure 4 – A graph of air velocity entering the lengthwise centerline of one of theopenings in the lower edge of the box shown in figure 1. The line is a plot of airvelocities along this centerline as predicted by the Simulation. The individual points(square, triangle, X) are velocities measured during the actual test trials.

APPLICATIONS OF SOLIDWORKS FLOW SIMULATION COMPUTATIONAL FLUID DYNAMICSSOFTWARE TO THE INVESTIGATION OF FIRESBy: John B. Holecek, P.E., CFEIPage 10Flow Simulation Example 2: Combustibles Close to a Fireplace HearthSeveral codes and standards address the use of combustible construction ina hearth extension of a fireplace or the proximity of combustible constructionto a fireplace. These include NFPA 211 Standard for Chimneys, Fireplaces,Vents, and Solid Fuel-Burning Appliances and the International ResidentialCode. These codes, and the instructions of many factory-built fireplaces,prohibit the use of combustibles like lumber in building a hearth extension. Afire that allegedly arose from such construction was the focus of a large testfireplace and chimney that was constructed and tested at the Warren facility.This same construction has since been modeled in SOLIDWORKS andsubjected to a CFD simulation using Flow Simulation.The test apparatus, shown in figure 6, consisted of a factory built modularwood burning masonry fireplace with a listed metal chimney. The fireplaceand chimney were mounted in an insulated surround. The face of thefireplace was covered with plywood, roofing felt and stucco to include forminga mantle. The hearth extension was made of dimensional lumber / plywoodcovered with roofing felt and stucco. Thermocouples were mounted in thesystem in various locations of interest. Of particular interest was theintersection of the hearth extension wood with the fireplace face. In the actualphysical test, the fire itself was supplied with an average of 9.1kg (20 pounds)per hour of seasoned dry hardwoods which equates to approximately 36,663Watts (125,000 BTUH). This continued for approximately 5 hours. The firebuilt a bed of coals that generally measured between 593-704 C (11001300 F) near its bottom layer which contacted the underlaying firebrick.Figure 7 shows the location of several thermocouples measuringtemperatures in the testing of the fireplace. These locations are along theleading top edge of the wood framing underlaying the stucco at the interfaceof the wood framing and the fireplace fire brick. This corner, given its closestproximity to the fire, would be the hottest area of the wood framing of thehearth extension.The CFD model was ran as an external analysis which includes theenvironment around the assembly. Flow was allowed into and out of theopenings of the fireplace and chimney under the influence of the buoyanteffects of gravity due to heat from the burning wood. Only the fireplace,chimney and hearth extension were modeled as the area of inquiry was the

APPLICATIONS OF SOLIDWORKS FLOW SIMULATION COMPUTATIONAL FLUID DYNAMICSSOFTWARE TO THE INVESTIGATION OF FIRESBy: John B. Holecek, P.E., CFEIPage 11temperatures in the hearth extension and the other surfaces such as themantel were not expected to influence the hearth extension temperature toany significant degree.Conduction within solids was allowed as was convective heat transfer to thewalls. Thermal properties for the materials used in the fireplace, chimney andhearth extension were specified including thermal conductivity, specific heatand density. Convective heat transfer coefficients were calculated using themethodology of Section 7-9 of JP Holman’s Heat Transfer Text.3 Theseranged from 2.1-5.8 W/m2K (0.37-1.02 BTU/HrFT2F) for the differing surfaces.External air temperatures were set at 4.4 C (40 F), which was the averagetemperature on the day of the testing. Radiation was included in this model,between the internal surfaces of the firebox, the “fire” and the top of thehearth extension. Each surface was modeled as a blackbody.Figure 6 – Flow Modeling Test Apparatus of a Wood Burning Fireplace with HearthExtension. The Corresponding Solid Model is Shown to the Right. The hearthextension cladding has been shown transparent to show the internal construction.The block inside the fireplace is the “fire”.As previously indicated, Flow Simulation does not directly model thecombustion process. However, solids and fluids can be assigned heatgeneration rates or temperatures that can be fixed or variable. For thesimulation, two basic methods were used to attempt modeling the fire. The

APPLICATIONS OF SOLIDWORKS FLOW SIMULATION COMPUTATIONAL FLUID DYNAMICSSOFTWARE TO THE INVESTIGATION OF FIRESBy: John B. Holecek, P.E., CFEIPage 12first method was to create a solid object of approximately the same size asthe combined burning logs and bed of coals and ascribe to that object aninternal volumetric heat generation rate equal to the average energy input of36,663 Watts (125,000 BTUH). The second method was to create a solidobject of the size of the bed of coals and ascribe to it the average 649 C(1200 F) temperature of the bed of coals as measured in the test.Several simulations were ran using the volumetric heat generation ratemethod. Generally, the method tended to overestimate the temperature of the“fire” when the modeled volume was of a physical size approximating thevisual volume of the burning logs and bed of coals. This is likely because itconfines the energy release occurring in the reacting flame into the smallervolume of the logs and coals. In short, the surface area of the modeled solidthrough which the specified energy must be emitted was so small as torequire a higher temperature difference to transfer the heat than occurs innormal combustion. If this volumetric heat generation method is used tomodel a fire, it would be necessary to have a sufficiently large volume suchthat the temperature of the volume drops to a normal flame temperature forthe actual situation at hand. Further work may show this method can besuccessfully employed at a sufficiently large volume.The second method employed was to fix the temperature of the volumerepresenting the coals and logs at the average 649 C (1200 F) temperatureof the bed of coals as measured in the test. A practical use of the simulationwas to determine if combustible construction in the hearth extension wasexposed to temperatures sufficient to result in ignition. Examination of thegeometry of the wood hearth structure construction makes it apparent thatconduction heat transfer from the adjacent fire brick and hearth extensionstucco covering would be the dominate form of heat transfer to theunderlaying combustible material. On this basis, for the specific purpose ofmodeling the temperature of the combustible hearth extension framing, theoverall heat generation rate of the fire, mostly lost as convective heating of airand radiant heat to the environment, was considered less of a significantinput as compared to the known temperature of the bed of coals. This fixedmethod appeared to have better promise as a means of modeling the fire forthis specific purpose.A comparison of the simulation data with temperatures recorded in the test isshown in figure 8. The solid lines that increase from the left are actualtemperatures recorded in the physical test, increasing with time. The values

APPLICATIONS OF SOLIDWORKS FLOW SIMULATION COMPUTATIONAL FLUID DYNAMICSSOFTWARE TO THE INVESTIGATION OF FIRESBy: John B. Holecek, P.E., CFEIPage 13can be seen to vary, which is to be expected as the fire was periodicallysupplied with logs. Thus, the energy input to the actual fire was periodic andnot uniform with time as would be the case for a gas fireplace. The horizontaldashed lines are steady state predicated values based on the simulation.These values are shown for the entire upper leading edge of the combustibleframing and a line 2.54cm (1 inch) back from the front edge.An examination of the graph of actual versus predicted temperatures (figure8) shows a reasonably close agreement between actual and predictedvalues. Considering the middle of the front edge, sensors 104 and 107, theaverage of sensor 104 was 327 C (621 F) and sensor 107 was 335 C(636 F) during the 2.5-hour period of leveled operation after the fire had builta bed of coals. This can be compared with the simulation predicted averagetemperature of 359 C (678 F) for the middle 30.5cm (12 inches) of theleading edge. Comparing these values, the agreement was within -5.3% and-3.8% respectively, calculated as the difference between actual and predictedvalues over the value of the actual average temperature in absolute units(degrees Kelvin). Moving back 2.54cm (1 inch) from the edge but still in themiddle, sensor 106 has an average value of 218 C (425 F) over the same2.5-hour period compared to the predicted value of 272 C (522 F), adifference of 11%.The variation between predicated and actualtemperatures is greater 2.54 cm from the edge than at the front edge.Figure 7: A view of thermocouple locations 104, 105 and 107 located against the top leading edge of thecombustible wood framing undelaying the stucco covering of the hearth extension. Thermocouples 102 and 106were one inch back from the leading edge. The photograph was taken prior to covering with stucco.

APPLICATIONS OF SOLIDWORKS FLOW SIMULATION COMPUTATIONAL FLUID DYNAMICSSOFTWARE TO THE INVESTIGATION OF FIRESBy: John B. Holecek, P.E., CFEIPage 14Figure 8: A graph of measured versus predicted wood temperatures. The dashed lines are predicted steady statetemperatures. The solid lines are measured temperatures during an actual test.SummaryIn the previous sections, SOLIDWORKS Flow Simulation has beenintroduced as a CFD program with the ability to assist in solving fireinvestigation problems. The basic process of using Flow Simulation has beendiscussed and two such applications to fire investigation have beendisclosed. In the first application, a method of modeling a fuel gas firedburner in Flow Simulation was demonstrated. Using that method, asimulation was conducted that modeled a test apparatus utilizing an actual LPgas fired burner. The simulation was shown to predict temperatures and flowrates that were reasonably close to the actual values. In the secondapplication, a method was shown that used a fixed elevated temperature solidto model a fire for the specific purpose of evaluating heat transfer to a hearthextension of combustible construction. This simulation yielded results thatwere reasonably close to the actual values.It has been the goal of this research to assist the reader with applyingSOLIDWORKS Flow Simulation to questions that arise in fire investigation.

APPLICATIONS OF SOLIDWORKS FLOW SIMULATION COMPUTATIONAL FLUID DYNAMICSSOFTWARE TO THE INVESTIGATION OF FIRESBy: John B. Holecek, P.E., CFEIPage 15Whereas these two simulations yielded reasonably close agreement betweenactual and predicted values, the extension of these methods to alternativesimulations should be undertaken with care. As with any use of a simulation,careful consideration should be given to defining the problem, selecting thecandidate model and model verification and validation.The use ofsimulations in fire investigation should be within the framework outlined inNFPA 921, which can be consulted for additional information.4Endnotes1. Ivanov, Trebunskikh, Platonovich: “Validation Methodology forModern CAD-Embedded CFD Code: from Fundamental Tests toIndustrial Benchmarks”. SolidWorks White Paper (2014)2. Technical Reference, SOLIDWORKS Flow Simulation 2017,Dassault Systemes (2017)3. Holman, JP, Heat Transfer, McGraw-H

Introduction Computational Fluid Dynamics (CFD) is a branch of applied science that utilizes computer numerical methods to solve problems of fluid flows and heat transfer. These problems are generally complex such that closed form . SOLIDWORKS Flow Simulation has integrated capabilities for all three steps. Additionally, SOLIDWORKS Flow .